From: "Christian König" <ckoenig.leichtzumerken-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
Subject: [PATCH 2/7] drm/amdgpu: rework shadow handling during PD clear
Date: Tue, 19 Feb 2019 14:40:58 +0100 [thread overview]
Message-ID: <20190219134103.1531-2-christian.koenig@amd.com> (raw)
In-Reply-To: <20190219134103.1531-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
This way we only deal with the real BO in here.
Signed-off-by: Christian König <christian.koenig@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 64 ++++++++++++++++----------
1 file changed, 39 insertions(+), 25 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 12d51d96491e..3c7b98a758c9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -788,38 +788,58 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
- goto error;
+ return r;
r = amdgpu_ttm_alloc_gart(&bo->tbo);
if (r)
return r;
+ if (bo->shadow) {
+ r = ttm_bo_validate(&bo->shadow->tbo, &bo->shadow->placement,
+ &ctx);
+ if (r)
+ return r;
+
+ r = amdgpu_ttm_alloc_gart(&bo->shadow->tbo);
+ if (r)
+ return r;
+
+ }
+
r = amdgpu_job_alloc_with_ib(adev, 64, &job);
if (r)
- goto error;
+ return r;
- addr = amdgpu_bo_gpu_offset(bo);
- if (ats_entries) {
- uint64_t ats_value;
+ while (1) {
+ addr = amdgpu_bo_gpu_offset(bo);
+ if (ats_entries) {
+ uint64_t ats_value;
- ats_value = AMDGPU_PTE_DEFAULT_ATC;
- if (level != AMDGPU_VM_PTB)
- ats_value |= AMDGPU_PDE_PTE;
+ ats_value = AMDGPU_PTE_DEFAULT_ATC;
+ if (level != AMDGPU_VM_PTB)
+ ats_value |= AMDGPU_PDE_PTE;
- amdgpu_vm_set_pte_pde(adev, &job->ibs[0], addr, 0,
- ats_entries, 0, ats_value);
- addr += ats_entries * 8;
- }
+ amdgpu_vm_set_pte_pde(adev, &job->ibs[0], addr, 0,
+ ats_entries, 0, ats_value);
+ addr += ats_entries * 8;
+ }
- if (entries) {
- uint64_t value = 0;
+ if (entries) {
+ uint64_t value = 0;
- /* Workaround for fault priority problem on GMC9 */
- if (level == AMDGPU_VM_PTB && adev->asic_type >= CHIP_VEGA10)
- value = AMDGPU_PTE_EXECUTABLE;
+ /* Workaround for fault priority problem on GMC9 */
+ if (level == AMDGPU_VM_PTB &&
+ adev->asic_type >= CHIP_VEGA10)
+ value = AMDGPU_PTE_EXECUTABLE;
+
+ amdgpu_vm_set_pte_pde(adev, &job->ibs[0], addr, 0,
+ entries, 0, value);
+ }
- amdgpu_vm_set_pte_pde(adev, &job->ibs[0], addr, 0,
- entries, 0, value);
+ if (bo->shadow)
+ bo = bo->shadow;
+ else
+ break;
}
amdgpu_ring_pad_ib(ring, &job->ibs[0]);
@@ -838,16 +858,10 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
amdgpu_bo_fence(bo, fence, true);
dma_fence_put(fence);
- if (bo->shadow)
- return amdgpu_vm_clear_bo(adev, vm, bo->shadow,
- level, pte_support_ats);
-
return 0;
error_free:
amdgpu_job_free(job);
-
-error:
return r;
}
--
2.17.1
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
next prev parent reply other threads:[~2019-02-19 13:40 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-19 13:40 [PATCH 1/7] drm/amdgpu: clear PDs/PTs only after initializing them Christian König
[not found] ` <20190219134103.1531-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2019-02-19 13:40 ` Christian König [this message]
[not found] ` <20190219134103.1531-2-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2019-02-20 0:07 ` [PATCH 2/7] drm/amdgpu: rework shadow handling during PD clear Kuehling, Felix
2019-02-22 15:51 ` Zeng, Oak
[not found] ` <BL0PR12MB258008B300413D2A0052EE5F807F0-b4cIHhjg/p/XzH18dTCKOgdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2019-02-22 16:34 ` Grodzovsky, Andrey
2019-02-19 13:40 ` [PATCH 3/7] drm/amdgpu: let amdgpu_vm_clear_bo figure out ats status Christian König
[not found] ` <20190219134103.1531-3-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2019-02-20 0:42 ` Kuehling, Felix
2019-02-19 13:41 ` [PATCH 4/7] drm/amdgpu: allocate VM PDs/PTs on demand Christian König
2019-02-19 13:41 ` [PATCH 5/7] drm/amdgpu: free " Christian König
2019-02-19 13:41 ` [PATCH 6/7] drm/amdgpu: drop the huge page flag Christian König
2019-02-19 13:41 ` [PATCH 7/7] drm/amdgpu: allow huge invalid mappings on GMC8 Christian König
2019-02-20 0:43 ` [PATCH 1/7] drm/amdgpu: clear PDs/PTs only after initializing them Kuehling, Felix
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190219134103.1531-2-christian.koenig@amd.com \
--to=ckoenig.leichtzumerken-re5jqeeqqe8avxtiumwx3w@public.gmane.org \
--cc=amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.